How this chapter was built

Every chapter runs through the same content pipeline, end to end:

• Book → corpus. The full 880-page Spectrum PDF was extracted (born-digital, no OCR needed) into a page-accurate text corpus.
• Corpus → spine. This chapter was broken into microtopics — the IDs that every question, note and test result hang from.
• Spine → chapter notes. A compressed revision layer shaped to what this chapter actually contains — tables, timelines, comparison points, trap alerts, and mnemonic devices.
• Spine → item bank. MCQs across the six real UPSC question forms, with trap-engineered wrong options and difficulty tags.
• Automated quality gate. Every question carries a citation to its printed page, and a verifier confirms each quoted passage appears verbatim in the book before the question enters the bank. A question that cannot cite its source is rejected, not shipped.

How the questions are benchmarked

The form mix leans the way real UPSC Prelims papers lean: statement-based and "how many of the above" questions dominate, with matching, chronology, direct-fact and assertion–reason making up the rest — calibrated against the 900-question, 2016–2026 PYQ corpus and its pattern report, not guessed. Wrong options are built from the trap patterns UPSC actually uses: negation flips, attribution swaps between similarly-named people or events, number exchanges, and on-paper-vs-in-practice gaps.

What the full build adds on top

Sectional and full-length mock engine with UPSC scoring (+2 / −0.667) and attempt-strategy tracking · weakness analytics per microtopic with page-precise reading prescriptions · parallel-form retests and improvement curves · spaced-repetition revision queue · CSAT module (same engine, 80Q, +2.5 / −0.833 profile).

Generated from the owner's copy of Spectrum (2019) for personal study use by one aspirant. Derived content must not be redistributed or used commercially.

Microtopic spine

Everything hangs from these IDs: questions, notes, test results, weakness scores, reading prescriptions.

Chapter Notes

Revision Sheet

The same tables, cards, trap alerts and mnemonics from Chapter Notes, pulled out on their own — for a fast last-look scan, no re-reading required.

Method & data model (for the full build)

Pipeline that produced this page

PDF → per-page corpus (corpus/spectrum-2019/pages/) → chapter map (book.json) → hand-curated, AI-drafted content (data.mjs) → verifier (verify.mjs: schema + verbatim passage check, 30/30 pass) → this packaged page (build.mjs). The study app will load the same JSON into MySQL; the app itself needs no AI at runtime.

Item record (what each question carries)

id · form · microtopic · difficulty · stem · options[4] · key · explanation · citation{page, verbatim quote} · traps[] — the citation is what makes wrong answers teachable: every explanation can say "Spectrum p. X" and mean it.

Question forms in the bank

Scoring used in the drill

GS Paper-I economics: correct +2.0, wrong −0.667, skipped 0. The sticky bar keeps a live net score so attempt discipline becomes visible early — this is the muscle the mock engine will train deliberately.